Quick answer: a wrong Primary Science answer is not one kind of failure. The student may have the wrong concept, miss a changed condition, misread a diagram, know the Science but lack the vocabulary to express it, confuse observation with inference, misunderstand investigation logic, or write an answer that hides a correct oral explanation. A useful tutor therefore classifies the first weak state before choosing the repair.

This page replaces an older generic “Primary Science Tuition” article. Its new job is diagnostic across Primary 3–6: when a Science answer is wrong, what exactly failed? That question sits beneath progression pages, readiness audits and PSLE preparation without competing with them.

MOE’s current Primary Science syllabus emphasises scientific knowledge, practices and values. That means useful diagnosis must look beyond keyword recall and include observation, representation, inquiry, evidence and communication.

MOE Primary Science Syllabus

Do not repair the final wrong sentence until you know the earliest scientific decision that failed.

The Nine Primary Science Error Classes

Error class	What it looks like	First repair
1. Concept	Scientific relationship itself is wrong	Rebuild model with examples/counterexamples
2. Condition	Answers familiar version, misses changed setup	Condition scan
3. Representation	Cannot translate diagram/table/setup into meaning	Switch representations
4. Vocabulary/language access	Concept known but question/terms block access	Clarify language and relationships
5. Evidence	Observation, inference and conclusion mixed	Bind claim to given evidence
6. Investigation	Variables/control/fair comparison misunderstood	Reconstruct experimental logic
7. Answer construction	Oral Science correct, written response weak	Represent mechanism explicitly
8. Execution/checking	Known Science lost through omission/timing	Question-specific checking/recovery
9. Transfer	Correct after teaching, fails changed question	Variation + delayed retest

1. Concept Error: The Scientific Model Is Wrong

A concept error occurs before answer technique. The learner’s internal model predicts the wrong relationship.

• Believes a process happens for the wrong reason.
• Confuses two related systems or functions.
• Uses a memorised keyword but connects it incorrectly.
• Cannot explain what changes when one condition changes.

Do not begin by polishing the sentence. Ask the student to explain orally, draw the system, predict a counterexample and compare two cases.

Concept Repair: Example → Counterexample → Changed Condition

1. State the current student model.
2. Use one clear example.
3. Use a case where the wrong model would predict incorrectly.
4. Redraw or re-explain the relationship.
5. Change one condition.
6. Ask the learner to predict again.

The model is stronger when it survives the changed condition.

2. Condition Error: The Student Answers the Question They Expected

Condition errors are common in familiar topics. Recognition of the chapter can cause premature recall.

System → what changed → what stayed the same → observation → relationship asked.

• Temperature changed but student uses the standard-condition answer.
• One part of a system is removed but the answer describes the intact system.
• An experiment reverses which group receives the treatment.
• A comparison asks “why less” while the student explains “why more”.

The repair is often reading discipline rather than more Science content.

3. Representation Error: The Science Is Hidden in the Diagram

Primary Science frequently moves among prose, diagrams, tables, cycles and experimental setups. A student may understand one representation but not another.

• diagram → explain in words;
• table → state trend/comparison;
• paragraph → draw causal relationship;
• setup → identify change, measure and controls;
• cycle → reconstruct sequence without picture.

Representation switching tests whether the Science belongs to the concept or merely to the worksheet surface.

4. Vocabulary / Language-Access Error

A learner can know the phenomenon but fail to access the question because one term, connector or sentence structure is unfamiliar.

Pattern	Diagnostic test
Unknown scientific term	Explain term in plain language
Question sentence misunderstood	Ask student to restate question
Oral explanation correct after rephrasing	Language access, not concept
Knows word definition but uses wrongly	Vocabulary relationship/collocation

The repair may be English support inside Science rather than reteaching the whole topic.

5. Evidence Error: Observation Is Not the Same as Explanation

Students often repeat what the table shows but fail to explain why, or make a conclusion stronger than the evidence permits.

State	Question
Observation	What was seen/measured?
Comparison	How did the groups differ?
Inference	What does this reasonably suggest?
Mechanism	What scientific model explains it?
Boundary	What would be too strong a conclusion?

A good Science answer keeps these states connected but not confused.

6. Investigation Error: The Comparison Logic Is Broken

Before technical labels, ask the plain-language questions:

• What are we deliberately changing?
• What are we observing or measuring?
• What should stay the same?
• Why should it stay the same?
• What result would support the idea?
• What can this experiment not prove?

If the student can name “independent variable” but cannot explain the fair comparison, the terminology has outrun the model.

Investigation Sub-Errors

• Changes more than one relevant condition.
• Measures the wrong outcome.
• Fails to keep starting conditions comparable.
• Chooses an unsuitable measuring method.
• Draws causal conclusions from insufficient evidence.
• Cannot explain why repeated or comparable measurements matter.

7. Answer-Construction Error: Correct Science, Invisible Science

This is one of the most important distinctions in Primary Science. Ask the child to explain the answer orally before judging the concept.

Oral	Written	Likely job
Wrong	Wrong	Concept/condition first
Correct	Vague	Answer construction
Correct	Keyword list	Causal linking
Correct	Too long/confused	Compression and relationship order

Condition → mechanism → resulting change → evidence/comparison.

This is a relationship spine, not a sentence template. The wording should remain natural to the student and the question.

8. Execution and Checking Error

Sometimes the Science is known but the paper performance fails.

• Misses “compare” and describes one item only.
• Forgets to use the given observation.
• Stops after naming a concept without explaining.
• Leaves an answer incomplete under time.
• Misreads an axis or table heading.
• Fails to return to a skipped question.

Checking should be question-specific: condition, evidence, units where relevant, comparison direction and final requested job.

9. Transfer Error: Correct Today, Wrong Tomorrow

A model answer can produce immediate recognition. Transfer asks whether the scientific operation survives after the surface changes.

1. Repair the original.
2. Close the model.
3. Reconstruct independently.
4. Change one condition or representation.
5. Ask again without naming the topic.
6. Return several days later.

Delayed changed-surface performance is stronger evidence of learning than immediate repetition.

One Wrong Answer Can Contain Several Error Classes

Suppose a student writes a weak explanation of an experiment. The final answer may contain:

• a missed changed condition;
• an inaccurate scientific term;
• correct observation but no mechanism;
• an overstrong conclusion;
• weak sentence construction.

Do not try to repair all five simultaneously. Find the earliest error that causes the later ones.

The First-Weak-State Routing Sequence

1. Can the student explain the concept orally? If no → concept.
2. Did the student notice the changed condition? If no → condition.
3. Can they interpret the diagram/table/setup? If no → representation.
4. Can they understand the question language? If no → language access.
5. Can they distinguish observation/evidence? If no → evidence.
6. Can they reconstruct the comparison? If no → investigation.
7. Can correct oral Science become writing? If no → answer construction.
8. Does the answer survive time/checking? If no → execution.
9. Does it survive a changed question later? If no → transfer.

Example: “My Child Knows the Keywords but Loses Marks”

Do not immediately add more keywords. Test:

• Can the child connect the keywords in a causal sentence?
• Can they draw the relationship?
• Can they use the idea on a different diagram?
• Can they explain why a tempting alternative is wrong?

The weakness may be concept structure or answer construction, not vocabulary volume.

Example: “My Child Understands Science but Cannot Answer Open-Ended Questions”

• Ask for oral explanation.
• Identify the question job.
• Mark the condition.
• State the mechanism.
• Bind it to the evidence.
• Write one concise answer.
• Change the surface and repeat.

If oral Science is correct, do not reteach the entire chapter.

Example: “My Child Is Careless in Science”

“Careless” is not a useful error class. Replace it with:

• condition omitted;
• comparison direction reversed;
• table heading ignored;
• evidence omitted;
• final requested quantity not answered;
• word copied incorrectly;
• checking routine absent.

Named errors can receive named prevention routines.

How the Taxonomy Changes From P3 to P6

Stage	Typical emphasis
P3	Observation, classification, vocabulary, basic concept relationships
P4	Condition reading, representations, evidence and simple investigation logic
P5	Integrated concepts, investigation reasoning, answer construction, transfer
P6	Mixed recognition, current-format execution, error budgets, verification and taper

The taxonomy stays stable; the complexity of the evidence changes.

How a 3-Pax Science Group Can Use One Taxonomy

eduKatePunggol’s current delivery model is maximum three students, typically 1.5 hours. One shared Science question can expose three different first weak states.

Shared question	Student A	Student B	Student C
Investigation explanation	Concept error	Condition/control error	Correct concept, answer-construction overreach

The curriculum is shared. The repair differs.

A Parent / Tutor Error Log

Question	Record
What was the final wrong answer?
What was the first weak state?
Which error class?
What is the smallest repair?
How will the surface change?
When will we retest?
Did hint size reduce?

What the Taxonomy Is Not

• It is not an official MOE marking taxonomy.
• It is not a grade predictor.
• It does not replace the syllabus.
• It does not diagnose medical or developmental conditions.
• It is not a reason to create extra worksheets for every category.

It is a teaching framework for turning a wrong answer into a more precise next move.

How This Connects to the Primary Science Route

For the broad progression, see How Primary Science Changes from P3 to P6. For current programmes, see Science Tuition at eduKatePunggol. For PSLE high-target planning, see Planning Toward AL1 in PSLE Science.

Responsible Claims

Classifying recurring Science errors can improve the precision of teaching, practice and feedback. It cannot guarantee a PSLE grade. Students may have more than one active error class, and the taxonomy should be revised when new evidence appears.

Frequently Asked Questions

Which Science error should be fixed first?

Usually the earliest recurring error that causes later failures. Near an exam, frequency, mark cost and realistic recoverability also matter.

Can a wrong answer be a language problem rather than a Science problem?

Yes. Ask for an oral explanation and rephrase the question. If the Science becomes correct, the first barrier may be language access or answer representation.

Should we keep an error notebook?

A small active error log can help. Avoid a giant archive of every mistake. Keep recurring causes and representative examples, then remove them when transfer is stable.

The Main Principle

Wrong answers are evidence. Classify them before you multiply them.

Check the concept. Check the condition. Translate the representation. Check the language. Separate observation from inference. Reconstruct the investigation. Make the mechanism visible. Verify the execution. Change the surface. Return later. The first weak state tells you what the next lesson should actually do.